INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS EXPERIMENTAL RESULTS KINETIC MODEL DISCUSSION REFERENCES

MONOAMINE OXIDASES (MAOs) purified from various tissues have been found to have rather high Michaelis constants for O₂ (K), generally >200 µM (18, 19, 24-26, 29, 33, 37, 44, 45) and well above the physiological intracellular O₂ concentrations in those tissues. This along with the observation that oxidative deamination of biogenic amines via MAOs can be a significant source of hydrogen peroxide (H₂O₂) involved in cell signaling (34, 41, 46) and oxidative injury (12, 14, 16, 51) suggests that the MAO O₂ sensitivity may be physiologically important. For example, the protective effect of MAO inhibition against central nervous system O₂ toxicity (51) and renal ischemia-reperfusion injury (16) has been attributed to the inhibition of the H₂O₂ produced during MAO-mediated monoamine metabolism. However, there is a large discrepancy between the K for MAO from the several studies carried out with purified MAO (18, 19, 24-26, 29, 33, 37, 44, 45) and the apparent K (12-34 µM) from the few studies carried out in cultured cells (26, 27). This suggests that in addition to possible species and/or tissue (31, 52) differences, the sensitivity for O₂ may depend on the cellular MAO environment (9, 40, 52).

The pulmonary endothelium is an important site of metabolism of certain monoamines present in the systemic venous return (15) and is subject to injury during high O₂ exposure (13). This brings into question the influence of PO₂ on pulmonary endothelial monoamine metabolism and a possible contribution of MAO-generated H₂O₂ to pulmonary hyperoxic injury. The discrepancy between results from purified enzymes and cell culture indicates that the O₂ dependency of MAO activity for cells residing in the intact organ cannot necessarily be predicted from studies on such reduced systems. Thus the objective of this study was to evaluate the O₂ dependency of the pulmonary endothelial MAO activity in intact lungs.

To this end, we developed a bolus injection multiple indicator dilution (MID) method for measuring MAO kinetics in the intact organ where the factors affecting substrate disposition can be more complex than in either cell culture or purified enzyme systems (2-6, 10, 22, 38). Thus a key aspect of the approach is the ability to separate the kinetics of substrate-tissue interactions (e.g., membrane transport, enzymatic metabolism, nonspecific plasma and tissue protein interactions) from each other and from the kinematics of organ perfusion (e.g., perfusion heterogeneity, transit time distribution) (4, 6, 10, 38). We measured the effect of PO₂ on the pulmonary venous effluent concentration of the MAO substrate -[ethyl-1-¹⁴C]phenylethylamine hydrochloride ([¹⁴C]PEA) and its [1-¹⁴C]phenylacetic acid ([¹⁴C]PAA) metabolite after the bolus injection of either [¹⁴C]PEA or [¹⁴C]PAA into the pulmonary artery of perfused rabbit lungs. The apparent K estimated with a kinetic model of the pulmonary disposition of [¹⁴C]PEA was ~18 µM and therefore much lower than that measured with purified enzymes (18, 19, 24-26, 29, 33, 37, 44, 45) and closer to that measured in cultured cells (26, 27).

Glossary

B	Site of sequestration of PEA within cells
CF(t)	3H2O effluent concentration versus time curve
Cin(t) = (q/F)hn(t)	Capillary input function
CR(t)	FITC-dextran (Dex) effluent concentration versus time curve
CT(t)	Indicator concentration versus time outflow curve for perfusion tubing system without the lung
hc(t)	Capillary transit time distribution
hn(t)	Noncapillary (arteries, veins, connecting tubing, and injection system) transit time distribution
K1 and K2	PEA and PAA endothelial surface equilibrium dissociation constants, respectively
k	Cell sequestration rate constant for PEA
[H2O][O2]k	MAO association rate constant
K	Michaelis constant for O2
kmet = ([H2O][O2]k)/3	Measure of the rate of PEA deamination by pargyline-sensitive MAO, where 3 = 1 + ([Pre]/Keq3)
kmet2	Measure of the rate of PEA metabolism via semicarbazide-sensitive form of MAO (SSMAO)
Keq2 and Keq1	PEA and PAA plasma protein equilibrium dissociation constants, respectively
Keq3 and Keq4	PEA and PAA equilibrium dissociation constants for nonspecific intracellular associations, respectively
kPAA = PS2/4Qe	Measure of PAA egress from the cells where 4 = 1 + ([Pre]/Keq4)
kPEA = PS1/3Qe	Measure of PEA egress from the cells where 3 = 1 + ([Pre]/Keq3)
kseq = k[B]Qe/3	Measure of the PEA sequestration rate within the lung tissue where 3 = 1 + ([Pre]/Keq3)
MAO	Monoamine oxidase
Pa and Pv	Arterial and venous pressures, respectively
PAA	Phenylacetic acid
PAAe(x,t)	(4Qe)[PAAe](x,t) where 3 = 1 + ([Pre]/Keq3)
PEA	Phenylethylamine
PEAB	PEA bound to site of sequestration within cells
PEAe(x,t)	(3Qe)[PEAe](x,t) where 3 = 1 + ([Pre]/Keq3)
PEAPre and PAAPre	PEA and PAA associated with nonspecific intracellular sites, respectively
PEAPrv and PAAPrv	PEA and PAA bound to plasma protein, respectively
[PAAc](x,t)	Vascular concentration of PAA at distance x from capillary inlet and time t
[PAAe](x,t)	Endothelial cell concentration of PAA at distance x from capillary inlet and time t
[PEAc](x,t)	Vascular concentration of PEA at distance x from capillary inlet and time t
[PEAe](x,t)	Endothelial cell concentration of PEA at distance x from capillary inlet and time t
Prv and Pre	Plasma and intracellular proteins, respectively
PS1 and PS2	Endothelial influx permeability-surface area products for PEA and PAA, respectively
q	Mass of the injected indicator
	Total organ flow
Qc	Capillary volume
Qe	Extravascular volume accessible to PEA and PAA
Qv	Pulmonary vascular volume
QW	Extravascular water volume
Q1 and Q2	Measures of the magnitude of the rapidly equilibrating nonspecific interactions of PEA and PAA, respectively, with luminal endothelial surface
[R](x,t)	Vascular concentration of the vascular reference indicator at distance x from the capillary inlet (x = 0) and time t
t	Time
W	Average linear flow velocity within Qc
x	Distance from the capillary inlet (x = 0)
Z1 and Z2	Nonspecific PEA and PAA surface binding sites, respectively

	EXPERIMENTAL METHODS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS EXPERIMENTAL RESULTS KINETIC MODEL DISCUSSION REFERENCES

Isolated Rabbit Lung Preparation

To produce a bolus injection, a solenoid-operated injection loop (3) was situated in the inflow tubing so that a 1.0-ml bolus could be introduced into the inflow stream without changing the flow or pressure. Just before the injection, the ventilator was stopped at end expiration and the venous outflow was directed into the sample tubes of a modified (3) Gilson Escargot fraction collector. One hundred 2-ml samples were collected at a sampling interval ranging from 0.3 to 2.4 s depending on the flow as described in Experimental Protocols.

After each experiment, the lungs were removed from the perfusion system, and additional bolus injections were made at the various flows studied, with the arterial and venous cannulas connected directly together. The data from these injections were used to obtain the concentration versus time curves [C_T(t)] and moments thereof (3, 6, 7) for the passage of the bolus through the tubing from injection to fraction collector in the absence of the lungs at each of the flows studied.

Bolus Composition

Sample Composition

The identities of the compounds in the venous effluent samples collected after [¹⁴C]PEA injection were established with thin-layer chromatography (TLC) on silica gel 60 TLC plates developed in an ethyl acetate-isopropanol-25% NH₄OH (50:35:10 vol/vol) mobile phase. With the use of authentic standards, the solute-to-solvent migration ratio values for PEA, PAA, and the intermediary metabolite phenylacetaldehyde were 0.46, 0.26, and 0.92, respectively. TLC analysis of selected peak venous outflow samples collected over the range of conditions studied demonstrated the presence of only [¹⁴C]PEA and [¹⁴C]PAA. After the identities of the ¹⁴C-labeled compounds in the venous effluent were established, [¹⁴C]PEA and [¹⁴C]PAA in each sample collected were separated with ion-exchange chromatography (43).

Bio-Rex 70 cation-exchange resin (200-400 mesh) was washed and equilibrated to pH 6.0 with 0.05 M sodium phosphate buffer. The resin was packed to a bed height of 1 cm in a plastic Poly-Prep column (Bio-Rad). The lyophilized 1.0-ml samples were redissolved in 2.0 ml of the pH 6.0 buffer before being passed through the columns. This was followed by two 1.0-ml water washes, and the total effluent from each column, which contained mostly [¹⁴C]PAA, was collected in a scintillation vial. The [¹⁴C]PEA was eluted from each column with two 2.0-ml aliquots of 0.25 M HCl, and the effluent was collected in a separate scintillation vial. The ¹⁴C counts in all samples were determined after the addition of 8 ml of Liquiscint (National Diagnostics, Atlanta, GA) on a Packard model 4530 liquid scintillation spectrometer. Recovery of ¹⁴C was >95%. The crossovers of PAA into PEA and of PEA into PAA were <0.5 and <5%, respectively, as measured with standards treated as the samples. These percentages include the inherent crossover of the ion-exchange separation procedure and any metabolism that might have occurred in postcollection samples (1). The venous effluent data measured after the bolus injection of [¹⁴C]PEA under the various experimental conditions described in Experimental Protocols were corrected for the crossover of PAA into PEA and of PEA into PAA.

Experimental Protocols

Flow. One experimental approach for separating the various processes affecting the pulmonary disposition of a given indicator is to vary the flow (2, 3, 6), which, in turn, varies the time the injected indicators are in contact with the pulmonary endothelium. To determine a useful range of flows in this context, a bolus containing FITC-Dex and either [¹⁴C]PEA (n = 7 lungs) or [¹⁴C]PAA (n = 1 lung) was injected with the flow set at 400, 200, 100, or 50 ml/min at outflow sampling intervals of 0.3, 0.6, 1.2, or 2.4 s, respectively. The [¹⁴C]PEA and [¹⁴C]PAA bolus injections were carried out in different lungs, and the flow sequence for these and subsequent experiments was randomized. The lungs were ventilated with the high O₂ gas mixture to maximize [¹⁴C]PEA metabolism. The effect of flow on the vascular volume was minimized by setting the mean pulmonary vascular pressure [(Pa + Pv)/2] at all flows to approximately equal that at 400 ml/min by adjusting the height of the venous outflow (2, 3, 6).

MAO inhibition. To evaluate the role of MAO in the pulmonary metabolism of PEA and to provide a positive control for evaluating the ability of the kinetic analysis (see KINETIC MODEL) to distinguish between changes in PEA uptake and metabolism, experiments with the MAO inhibitors pargyline and semicarbazide (21, 43) were carried out by perfusing the lungs with perfusate containing 20 µM pargyline and 1.0 mM semicarbazide for 5 min before the [¹⁴C]PEA (n = 4 lungs) or [¹⁴C]PAA (n = 1 lung) bolus injections at 400 and 50 ml/min and high PO₂. Pargyline (20 µM) and semicarbazide (1.0 mM) were also included in the injected boluses. The pargyline and semicarbazide concentrations were chosen because they had been previously shown to completely inhibit [¹⁴C]PEA metabolism by perfused rabbit lungs (21, 43).

Varying PO₂. The O₂ dependency of [¹⁴C]PEA and [¹⁴C]PAA pulmonary disposition was measured by injecting boluses containing either [¹⁴C]PEA or [¹⁴C]PAA at one or more of the O₂ levels at 400 and 50 ml/min. During the transition to the low PO₂, there was a transient increase in perfusion pressure, reflecting hypoxic vasoconstriction. This constriction had dissipated by the time the low PO₂ had been reached as previously described (39).

Additional Experiments

Drugs and Isotopes

	EXPERIMENTAL RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS EXPERIMENTAL RESULTS KINETIC MODEL DISCUSSION REFERENCES

Figure 1 exemplifies the venous [¹⁴C]PEA and [¹⁴C]PAA concentration curves after pulmonary arterial injection of either [¹⁴C]PEA or [¹⁴C]PAA over the range of perfusate flows indicated. Figure 1, top, demonstrates that [¹⁴C]PEA is extensively extracted by the lungs. For instance, at 400 ml/min, which results in a capillary mean transit time of <1 s (5), >70% of the injected [¹⁴C]PEA was extracted after a single pass through the pulmonary circulation. The ability of the lungs to metabolize [¹⁴C]PEA is demonstrated by the appearance of [¹⁴C]PAA in the venous effluent after the injection of [¹⁴C]PEA. At 400 ml/min, the amount of [¹⁴C]PAA in the venous effluent was relatively low. Decreasing the flow increased both [¹⁴C]PEA uptake and the appearance of [¹⁴C]PAA. At 50 ml/min, >77% of the injected [¹⁴C]PEA was recovered as [¹⁴C]PAA in the venous effluent compared with ~21% at 400 ml/min as indicated in Table 1 and Fig. 1.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Pulmonary venous concentration (normalized to amount of injected indicator) vs. time curves for fluorescein isothiocyanate-dextran (FITC-Dex), [14C]phenylethylamine ([14C]PEA), and [14C]phenylacetic acid ([14C]PAA) after the bolus injection of FITC-Dex and either [14C]PEA (A-D) or [14C]PAA (E-H) into the pulmonary artery of an isolated rabbit lung (separate lungs for [14C]PEA and [14C]PAA injection) perfused at the indicated flows. The lungs were ventilated with the high O2 gas mixture. See text for model fits to data.

The [¹⁴C]PAA curves obtained after [¹⁴C]PAA injection provide information about the pulmonary disposition of [¹⁴C]PAA independent of [¹⁴C]PEA metabolism, which is required for subsequent kinetic analysis. It is clear from the data in Fig. 1, bottom, that the rate of [¹⁴C]PAA uptake is much slower than that of [¹⁴C]PEA.

Treatment with both pargyline and semicarbazide decreased the [¹⁴C]PAA concentration in the venous effluent after [¹⁴C]PEA injection to undetectable levels at both high and low flows (Fig. 2 and Table 1). Figure 3C shows that treatment with only pargyline decreased the fractional recovery of [¹⁴C]PAA in the collected venous effluent samples after [¹⁴C]PEA bolus injection at 50 ml/min by ~ 80%. On the other hand, any effect of treatment with semicarbazide alone was undetectable (Fig. 3B).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Pulmonary venous concentration (normalized to amount of injected indicator) vs. time curves for FITC-Dex, [14C]PEA, and [14C]PAA after the bolus injection of these indicators into the pulmonary artery of a rabbit lung perfused at 400 (A) and 50 (B) ml/min after the lung had been treated with the MAO inhibitors pargyline (20 µM) plus semicarbazide (1 mM). See text for model fits to data.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Pulmonary venous concentration (normalized to amount of injected indicator) vs. time curves for FITC-Dex, [14C]PEA, and [14C]PAA after the bolus injection of these indicators into the pulmonary artery of a rabbit lung perfused at 50 ml/min before (control; A) and after the lung was treated with either 1 mM semicarbazide (B) or 20 µM pargyline (C). Separate lungs were used for pargyline and semicarbazide, but only one set of pretreatment control curves is shown because the control data from the 2 lungs are virtually superimposable. See text for model fits to data.

The O₂ dependency of [¹⁴C]PEA metabolism in the intact lung is exemplified in Fig. 4. No O₂ effect is detectable in the range of PO₂ values from 518 to 106 Torr. However, at a PO₂ of ~16 Torr, [¹⁴C]PAA concentrations in the venous effluent after [¹⁴C]PEA injection were lower (most notably at 50 ml/min) than those at a high PO₂ (Fig. 4).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Pulmonary venous concentration (normalized to amount of injected indicator) vs. time curves for FITC-Dex, [14C]PEA, and [14C]PAA after the bolus injection of these indicators into the pulmonary artery of an isolated perfused rabbit lung perfused at the indicated flows after the average of the arterial inflow and venous outflow PO2 had reached 518 Torr (A), 106 Torr (B), or 16 Torr (C). All the data were obtained from the same lung. See text for model fits to data.

Neither the MAO inhibitors nor PO₂ had a significant effect on the [¹⁴C]PAA outflow curves after [¹⁴C]PAA bolus injection (see Estimation of Model Parameters). Thus only the PAA outflow curves measured after [¹⁴C]PAA bolus injection in lungs ventilated with the high PO₂ gas mixture are shown (Fig. 1).

The fractions of injected [¹⁴C]PEA recovered in the collected venous effluent samples as [¹⁴C]PEA or [¹⁴C]PAA are given in Table 1 for each of the experimental conditions studied. The fractions of injected [¹⁴C]PAA and FITC-Dex recovered were 98.5 ± 2.7 (SD) and 95.2 ± 1.6%, respectively.

The Pa and Pv values at 400 and 50 ml/min under the various experimental conditions studied are given in Table 2. The pulmonary vascular volume (Q_v) and extravascular water volume (Q_W) calculated from the FITC-Dex, C_R(t), ³H₂O, C_F(t), and tubing C_T(t) outflow curves (7) under the various experimental conditions studied were 8.6 ± 1.3 (SD) and 6.8 ± 1.4 ml, respectively, and were not significantly affected by experimental condition or flow.

	KINETIC MODEL

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS EXPERIMENTAL RESULTS KINETIC MODEL DISCUSSION REFERENCES

Reactions

(a)

(b)

Single-Capillary Element

(1)

(2)

(3)

(4)

(5)

The identifiable model parameters are k_met (s¹), which is the measure of the rate of PEA deamination by MAO; k_seq (s¹), which is the measure of the PEA sequestration rate within the lung tissue; PS₁ (ml/s) and PS₂ (ml/s), which are the endothelial permeability-surface area products for PEA and PAA, respectively; k_PEA (s¹) and k_PAA (s¹), which are measures of the respective rate of PEA and PAA egress from the cells; and the virtual volumes Q₁ (ml) and Q₂ (ml), which are measures of the magnitude of the rapidly equilibrating cell surface interactions of PEA and PAA, respectively. For PAA injections, Eqs. 2-5 reduce to Eqs. 4 and 5 with PEA_e set to zero, and the number of identifiable parameters reduces to three, namely PS₂ (ml/s), k_PAA (s¹), and Q₂ (ml). The values of ₁ and ₂ were set at 1.14 and 9.1, respectively, based on the measured PEA-BSA and PAA-BSA binding.

Whole Organ

Estimation of the kinetic model parameters described in Estimation of Model Parameters involved numerically [finite difference method (4)] solving Eqs. 1-5 for the appropriate boundary conditions at each iteration of a Levenberg-Marquardt optimization routine (35). The time step was chosen by successively halving an initial time step until the coefficients of variation between the solutions of Eqs. 1-5 at successive time steps was <2%.

Estimation of Model Parameters

The first step in the parameter estimation procedure was to utilize the [¹⁴C]PAA data after [¹⁴C]PAA injection to estimate the PAA model parameters, namely PS₂, k_PAA, and Q₂, independently of PEA uptake and metabolism to PAA. This was accomplished by simultaneously fitting the solutions of Eqs. 1, 4, and 5 with the initial conditions [R](x,0) = [PAA_c](x,0) = 0 and PAA_e(x,0) = 0 and the boundary conditions [R](0,t) = C_in(t); [PAA_c](0,t) = (1/₂) C_in(t); and PAA_e(0,t) = 0 to the [¹⁴C]PAA concentration versus time data measured after [¹⁴C]PAA bolus injections at 400 and 50 ml/min. C_in(t) = (q/)h_n(t) is the capillary input concentration curve (2-6), where h_n(t) is the noncapillary (arteries, veins, connecting tubing, and the injection system) transit time distribution, and q and are the mass of the injected indicator and total flow through the organ, respectively. C_in(t) is related to the vascular reference indicator curve C_R(t) and the h_c(t) by the convolution relationship C_R(t) = C_in(t)*h_c(t) as previously described (2-6). Table 3 shows the estimates of the PAA model parameters and measures of precision of these estimates, namely the 95% confidence intervals and the correlation matrix (3, 30). To determine whether the values of the PAA model parameters were affected by any of the experimental conditions, we compared the model fit obtained using the parameter values estimated from each individual experiment with the fit obtained using the mean set of PAA model parameters values estimated from all the PAA experiments given in Table 3. The F ratios (36) indicated that the fits to the individual data sets using the mean parameters were not significantly worse than those using individual parameters estimated from each data set. Thus it is concluded that any effects of the different experimental conditions were not detectable, and the kinetic model parameters descriptive of the pulmonary disposition of PAA were set to the mean values of 0.27 ml/ml of vascular volume, 0.063 s¹, and 0.50 ml · s¹ · ml¹ of vascular volume for PS₂/Q_v, k_PAA, and Q₂/Q_v, respectively. The normalization to the Q_v measured for each lung is to accommodate small differences in lung sizes.

Knowing the PAA kinetic parameters, the parameters descriptive of PEA disposition, namely k_met (s¹), k_seq (s¹), PS₁ (ml/s), k_PEA (1/s), and Q₁ (ml) were estimated by fitting the model to the [¹⁴C]PEA and [¹⁴C]PAA data obtained after [¹⁴C]PEA injection. This was accomplished by simultaneously fitting the solutions of Eqs. 1-5 with the initial conditions [R](x,0) = [PEA_c](x,0) = [PAA_c](x,0) = 0 and PEA_e(x,0) = PAA_e(x,0) = 0 and the boundary conditions [R](0,t) = C_in(t); [PEA_c](0,t) = (1/₁)C_in(t); [PAA_c](0,t) = 0; and PEA_e(0,t) = PAA_e(0,t) = 0 to the [¹⁴C]PEA and [¹⁴C]PAA data after [¹⁴C]PEA bolus injections at 400 and 50 ml/min. With this parameter estimation approach, the effects of MAO inhibitors and PO₂ on the pulmonary disposition of PEA were evaluated.

Table 4 shows the estimated values of the PEA model parameters and measures of precision of these estimates (3, 30) under the various experimental conditions studied. The resulting model fit is exemplified in Figs. 2 and 4. The estimated values of the PEA model parameters are also shown in Table 5, where the extensive parameters PS₁ and Q₁ were normalized to the Q_v to account for small differences in lung sizes. With pargyline and semicarbazide treatment, k_met became undetectable with little effect on PS₁. The effects on k_seq and Q₁ were small but significant. In the range from 518 to 106 Torr, PO₂ had no significant effect on the PEA kinetic parameters (Table 5). However, at PO₂ = 16 Torr, the estimated value of k_met was significantly smaller than that estimated at the higher PO₂ levels. The effect of PO₂ = 16 Torr on k_seq was also significant. The apparent K for MAO in the intact lung was estimated from k_met = [(k_met)_maxPO₂]/(PO₂ + K) (Fig. 5), where (k_met)_max is the maximum k_met. The value was 17.0 Torr (18.2 µM).

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Rate of PEA deamination by pargyline-sensitive monoamine oxidase (kmet) at the 3 O2 levels studied. Solid line superimposed on data is the fit of kmet = (kmet)maxPO2/PO2 + K, where (kmet)max is the maximum kmet and K is the Michaelis constant for O2. Values are means ± SE. The apparent K is 17.0 Torr (18.2 µM).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS EXPERIMENTAL RESULTS KINETIC MODEL DISCUSSION REFERENCES

The results indicate that the apparent K for MAO in the intact lung is significantly smaller than that estimated from studies carried out with purified MAO (18, 19, 24-26, 29, 33, 37, 44, 45) and closer to that estimated from studies carried out in cultured cardiac myocytes (27) and hepatocytes (26). The value is well below normal alveolar PO₂, suggesting a minimal influence of PO₂ on lung MAO activity. The results are consistent with differences in the activities and specificities of enzymes in general (8, 49) and of MAO in particular (9, 49) in their in situ cellular environments versus the purified enzymes. Although the estimated value of K is much smaller than most of those estimated for purified MAO and probably not relevant in the hyperoxic range, it is still in a range high enough for sensing changes in O₂ in the hypoxic range. Furthermore, the results do not preclude an important role for MAO-generated H₂O₂ in the face of increased plasma concentrations of biogenic amines over any range of PO₂ values (23, 51).

The O₂ concentrations used to estimate K (Fig. 5) were measured in the perfusate. These values may be higher than the O₂ concentrations near the outer mitochondrial membrane site of MAO under the assumption that the mitochondria are sinks for O₂. However, in the lungs, where the rate of metabolism is low compared with the rate of gas transport across the alveolar capillary barrier and where diffusion distances are small, the expectation is that the local PO₂ is not much lower than the perfusate PO₂ (27). Thus although the estimated K in this sense represents an upper bound on the actual K, the difference is probably quite small and in the direction of increasing the discrepancy between intact lungs and purified enzymes. Previous studies (19, 27) on the O₂ dependency of the MAO reaction have demonstrated that the K is also somewhat sensitive to the monoamine substrate concentration, consistent with the nature of the MAO reaction mechanism (19, 27). In the present study, the estimate is for tracer PEA concentration.

One aspect of studies on intact organs and cells that distinguishes them from studies on purified enzymes is that the rate of entry into the cells needs to be accounted for. In the above model, PEA uptake is represented by a linear transport mechanism (9) having a permeability-surface area product (PS₁). Previous investigations (21, 48) have concluded that this mechanism is passive diffusion. Given the relatively low lipid solubility of PEA, its extensive pulmonary uptake via passive diffusion is somewhat surprising (48). The data in Figs. 1-4 and the relative rates of uptake and metabolism indicate that the intracellular concentration of PEA during bolus passage became much greater than the vascular concentration. This would be consistent with a large PEA tissue-to-perfusate partition coefficient, which in the model would be a large ₃Q_e-to-₁Q_c ratio. Alternatively, if an active uptake mechanism were involved, the model representation would be a larger permeability-surface area product for PEA uptake (PS₁) than for egress, the latter being lumped with other processes in the group parameter k_PEA. These two effects are not separable with the data obtained in the present study. However, with additional experimental protocols, the kinetic model and the MID method developed in the present study may be useful for further evaluation of the underlying PEA uptake mechanism.

Another difference between studies with intact organs or cells and studies with purified enzymes is that in intact organs and cells, competing processes such as substrate interactions with perfusate constituents (e.g., plasma proteins) and with any number of cellular and/or tissue constituents can also affect substrate availability by altering its partitioning within the cells and between the medium (perfusate) and the cells. These factors are represented in the model by Q₁ and Q₂ for nonspecific cell surface and/or tissue interactions, ₁ and ₂ for perfusate albumin interactions, and ₃, ₄, and k_seq for intracellular interactions.

The major difference between intact organs and isolated cells is that in the intact organ, access to the cells is generally via the vascular system, which further complicates the evaluation of intracellular functions. For example, changing the flow by itself has a substantial effect on the extracted fraction of PEA (Fig. 1), and at a given flow, the overall extraction fraction is determined by contributions from the capillary pathways with different flows and/or transit times (4, 5, 10) and hence different extraction fractions. This is taken into account in the model by allowing for longitudinal spatial variations in the concentration of the injected indicators within a given capillary element (Eqs. 1-5), by representing the organ by parallel capillary elements with different transit times (4), and by weighing the contributions of these capillary elements according to h_c(t) as previously discussed in more detail (4).

Previous studies (21, 43) have concluded that the endothelium is the main site for the MAO responsible for the oxidative deamination of PEA in the lung. Three forms of MAO have been identified in the rabbit lung (20, 21, 42, 43), including MAO-A, MAO-B, and a semicarbazide-sensitive form (SSMAO), with distinct subcellular localizations and substrate and/or inhibitor affinities. MAO-A and MAO-B are both located on the outer mitochondrial membrane (20, 21, 32, 42, 43), but MAO-A has a higher affinity for serotonin and norepinephrine and is selectively inhibited by clorgyline (21, 43), whereas MAO-B has a higher affinity for PEA and is more sensitive to inhibition by pargyline (21, 43). PEA is also a substrate for SSMAO, which is thought to be located on the plasma membrane (11, 17, 21, 43, 50). Roth and Gillis (43) and Gillis and Roth (21) found that treatment with pargyline followed by the addition of semicarbazide reduced PEA metabolism by 70 and ~100%, respectively. This is consistent with the results in the present study where treatment with pargyline alone decreased the fractional recovery of [¹⁴C]PAA in the collected venous effluent samples after a [¹⁴C]PEA bolus injection at 50 ml/min by ~80% (Fig. 3C) and that treatment with both pargyline and semicarbazide was needed for full inhibition (Fig. 2). However, the lack of effect of semicarbazide alone on PEA metabolism (Fig. 3B) has some interesting implications as indicated below.

In the kinetic model represented by Eqs. 2-5, there is no explicit accommodation for two types of MAO. To evaluate the possible implications of this simplification, we modified the kinetic model represented by Eqs. 1-5 to allow for PEA metabolism via both luminal surface (SSMAO) and intracellular MAO (MAO-B). This was accomplished by substituting Eqs. 2 and 5 with Eqs. 6 and 7

(6)

(7)

where k_met2 (ml/s) is a measure of the rate of PEA metabolism by SSMAO. The value of k_met2 was estimated by fitting the solution of Eqs. 3 and 5-7 to the [¹⁴C]PEA and [¹⁴C]PAA data measured after the bolus injection of [¹⁴C]PEA at 50 ml/min into the pulmonary artery of the lung treated with pargyline alone (Fig. 3C). The identifiable parameters were k_met2 and k_seq, with k_met set to zero and the other parameters set to the mean values in Table 3 (PO₂ = 518 Torr). The estimated value of k_met2 was 0.31 ml/s, which is much smaller than the ~16 ml/s estimated for PEA uptake (PS₁). This result indicates that as long as there is significant PEA uptake and sufficient PEA metabolism by pargyline-sensitive MAO, the data will be insensitive to the contribution of SSMAO to PEA metabolism. That explains why semicarbazide alone did not have a detectable effect on PEA metabolism (Fig. 3B). It also indicates that the apparent K estimated is for pargyline-sensitive MAO because the high rate of PEA uptake over the range of PO₂ studied (16-518 Torr) makes the SSMAO contribution negligible under the study conditions.

The kinetic model represented by Eqs. 2-5 allows for the previously observed PEA sequestration within the cells (21, 43, 47). Table 1 shows that treatment with MAO inhibitors significantly decreased the total amount of ¹⁴C recovered in the venous effluent samples collected after a [¹⁴C]PEA bolus injection at 50 ml/min (47). This decrease in total recovery reveals competition between the two intracellular processes, namely [¹⁴C]PEA metabolism and sequestration. The normal rate of [¹⁴C]PEA metabolism (k_met) is faster than that of [¹⁴C]PEA sequestration (k_seq) as shown in Table 5. Thus, in the absence of MAO inhibitors, most of the [¹⁴C]PEA extracted by the pulmonary endothelium was metabolized to [¹⁴C]PAA and ultimately returned to the perfusate (Fig. 1C). After treatment with MAO inhibitors, metabolism was no longer competing with sequestration for [¹⁴C]PEA, and hence a larger fraction of the PEA taken up was sequestrated and a smaller fraction of the injected ¹⁴C was recovered in the venous effluent. This competition between intracellular PEA metabolism and sequestration was not detectable at 400 ml/min (Table 1) because at this flow, less time was available for either [¹⁴C]PEA sequestration or metabolism than at the lower flow. Thus a larger fraction of the extracted PEA returned to the perfusate. This observation further demonstrates the utility of varying the flow for revealing competing parallel processes.

To help put this flow dependency (2, 3, 6) in perspective, the normalized sensitivity function S(t) (6) obtained for the PEA model parameters estimated from the [¹⁴C]PEA and [¹⁴C]PAA data after [¹⁴C]PEA injections at 400 and 50 ml/min are shown in Fig. 6. For the ith model parameter _i, S_i(t) = C(t)/_i, where C(t) is the calculated PEA or PAA indicator effluent concentration. The sensitivity function S_i(t) was approximated by the change in C(t) resulting from a 1% change in _i divided by the change in _i (6). Multiplying S_i(t) by the value of the parameter estimate, _i, provides an indication of the relative contribution of the parameter to the model fit to the data at a given time (6). Comparison of these normalized sensitivity functions reveals the extent and the time epoch to which the optimized model parameters make their contributions to the model fit. The shapes of the sensitivity function relative to each other reveal how independent the contributions of the individual parameters are to the model fit. For example, at 400 ml/min, the dominant role of PEA uptake in the model fit is revealed by the sensitivity function of PS₁ (Fig. 6A). The contribution of k_met at 400 ml/min is small relative to that for PS₁ (Fig. 6A). At 50 ml/min, on the other hand, k_met plays a dominant role, whereas the contribution of PS₁ is relatively small (Fig. 6D). Thus fitting the model to the data at both flows reduces the correlation between model parameters by increasing the extent of their contributions to the model fit to the data and by extending and segregating the time epochs over which they make their contributions.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Sensitivity fun